Active damping control within a supervisory control structure

The present application generally relates to hybrid electric vehicles and, more particularly, to a system and method for active damping control in a hybrid vehicle.

A hybrid electric vehicle includes an internal combustion engine (ICE), at least one high-voltage battery system and at least one electrified drive module having an electric motor and associated electric drive gearbox assembly. The ICE and electric motor(s) can be generally referred to as prime movers. To effectively manage the efficiencies of these prime movers, they require to be connected and disconnected to the driveline. Due to the complexity of the multiple paths, the driveline tends to have additional lash in the torque path as compared to conventional ICE drivelines. When inherent lash and stiffness of the driveline are not considered, the resulting command can lead to vibrations and oscillations in the powertrain, reducing performance and comfort. This lash needs to be controlled. Accordingly, while such hybrid powertrains do work well for their intended purpose, there is a desire for improvement in the relevant art.

According to one example aspect of the invention, an electrified powertrain that generates and transfers drive torque to a driveline of a hybrid electric vehicle includes a first electric motor, a second electric motor and a controller. The first electric motor includes a first electric motor output. The second electric motor includes a second electric motor output. The controller is configured to: receive a driver torque request; determine an open loop motor torque command based on the driver torque request; and determine a shaped torque command based on the open loop motor torque command, including identifying a lash zone, requesting motor torque from at least one of the first and second electric motors to cross the lash zone, and, subsequent to crossing the lash zone, providing motor torque to meet the driver torque request.

In some implementations, the controller is further configured to perform closed loop active electric motor damping of at least one of the first and second electric motors.

In some implementations, the controller further comprises a supervisory motor torque determination module that (i) receives a driver demand shaping signal based on the driver torque request and (ii) an engine torque signal.

In some implementations, the supervisory motor torque determination module further comprises an active damping controller that (i) receives the open loop motor torque command and (ii) outputs the shaped torque command to at least one of the first and second electric motors.

In additional features, the controller further includes a motor controller including an active electric motor dampening (AEMD) module that receives the shaped torque command and outputs modified torque actuation signals to the respective first and second electric motors.

In additional features, the motor controller implements a closed loop speed control that mitigates residual oscillations in the electrified powertrain.

In additional features, the electrified powertrain further comprises a supervisory motor torque determination module implements a linear quadratic integral (LQI)-based compensator that provides the shaped torque command.

In additional features, the electrified powertrain further comprises an internal combustion engine (ICE).

In additional features, a method for controlling an electrified powertrain that generates and transfers drive torque to at least one of a first and second electric motor in a driveline of a hybrid electric vehicle in provided. The method includes: receiving a driver torque request; determining an open loop motor torque command based on the driver torque request; determining a shaped torque command based on the open loop motor torque command, including identifying a lash zone; and requesting motor torque from at least one of the first and second electric motors to cross the lash zone, and, subsequent to crossing the lash zone, providing motor torque to meet the driver torque request.

In additional features, the method includes performing closed loop active electric motor damping of at least one of the first and second electric motors

In additional features, the method includes receiving, at a supervisory motor torque determination module, (i) a driver demand shaping signal based on the driver torque request and (ii) an engine torque signal.

In additional features, the method includes receiving, at an active damping controller, the open loop motor torque command; and outputting the shaped torque command to at least one of the first and second electric motors.

In additional features, the method includes: receiving, at a motor controller including an active electric motor dampening (AEMD) module, the shaped torque command; and outputting modified torque actuation signals to the respective first and second electric motors.

In additional features, the method includes implementing a closed loop speed control that mitigates residual oscillations in the electrified powertrain.

In additional features, the method includes implementing, at a supervisory motor torque determination module, a linear quadratic integral (LQI)-based compensator that provides the shaped torque command.

Further areas of applicability of the teachings of the present application will become apparent from the detailed description, claims and the drawings provided hereinafter, wherein like reference numerals refer to like features throughout the several views of the drawings. It should be understood that the detailed description, including disclosed embodiments and drawings referenced therein, are merely exemplary in nature intended for purposes of illustration only and are not intended to limit the scope of the present disclosure, its application or uses. Thus, variations that do not depart from the gist of the present application are intended to be within the scope of the present application.

As mentioned above, it is desirable to effectively manage the efficiencies of prime movers that are repeatedly connected and disconnected to the vehicle driveline in a hybrid powertrain. Due to the complexity of the multiple paths, the driveline tends to have additional lash in the torque path as compared to conventional ICE drivelines.

The present disclosure provides a novel system and method for controlling and actuating vibrations and torque noises in hybrid and electric vehicles due to flexible shafts and backlash within the drivetrain components. In particular, the present disclosure provides a model-based control technique based on a linear quadratic gaussian (LQG) formulation. The model-based control includes a linear-switching model of the driveline where the states and formulations are properly chosen to facilitate calibration of the control strategy. A rigid model is used to generate references when in contact. A physic-based reference logic is used to generate references when in the lash zone. A linear Kalman Filter is used to estimate the state of the system (contact, no-contact). A gain scheduling approach wherein the gains change depending upon the state of the system is used. The states of the system can include all parts in contact, some parts in lash, all parts in lash, etc. A linear Quadratic Integral (LQI) controller with a gain scheduling approach is provided. The LQI controller takes the estimated states, compares them to the reference states, and generates torque commands to prevent and compensate for any oscillations/vibrations within the system.

The supervisory control described herein optimizes multiple actuators and minimizes computational complexity. The control method simplifies the powertrain down to the minimum states it requires for optimization purposes. This results in the control not accounting for non-linearities and stiffness, which results in the generation of the optimum torque commands with a minimum computational burden. A model-based controller of a hybrid powertrain is disclosed that, in response to a driver-demanded reference trajectory, appropriately commands the electric machines as actuators, while accounting for the complex dynamics of the entire system plant. These dynamics encompass additional system states when considering stiffness and damping of the powertrain and incorporate nonlinearities, such as backlash, which introduces challenges into the control process.

Traditional torque control methodologies often employ torque shaping techniques that involve the use of nonlinear filters to modify the reference torque desired, especially in response to the lash zone, a region associated with backlash. Such filters conventionally introduces a deliberate delay in torque commands as they traverse the lash zone, releasing torque aggressively once the zone is traversed (explained in greater detail herein with respect to). Such an approach, derived from the context of internal combustion engines, may not be ideal for electric vehicles, which lack the inherent delays associated with the combustion process.

Referring now to, a functional block diagram of an example hybrid electric vehicle(also referred to herein as “vehicle”) according to the principles of the present application is illustrated. The vehicleincludes an electrified powertrainhaving an electrified drive module (EDM)configured to generate and transfer drive torque to a drivelinefor vehicle propulsion. The EDMgenerally includes one or more electric drive units or motors(e.g., electric traction motors), an electric drive gearbox assembly or transmission, and power electronics including a power inverter module (PIM). As will become appreciated herein, the exemplary powertrainincludes a first electric motorA, a second electric motorB, and a third electric motorC (). It is appreciated that while the example discussed herein provides three electric motors, the teachings are equally adapted for electric motor combinations less than or more than three. The transmissionis configured as a four speed EVT.

The electric motorsare connected via the PIMto a high voltage battery systemfor powering the electric motors. The battery systemis selectively connectable (e.g., by the driver) to an external charging system(also referred to herein as “charger”) for charging of the battery system. The battery systemincludes at least one battery pack assembly. The electrified powertrainis a hybrid powertrain that additionally includes an internal combustion engine (ICE). As will be described herein, the electric motorsand the ICEcooperate to selectively connect and disconnect with the drivelineto provide drive torque to drive wheels.

A vehicle control systemincludes a controllerthat can provide various inputs to the EDMincluding torque requests based on signals received from a driver interface. In examples, the driver interfacecan include a drive input device, e.g., an accelerator pedal, for providing a driver input, e.g., a torque request, to the controllerand ultimately the EDM. The driver interfacecan further include a human machine interface (HMI)for displaying driver information and receiving driver requests. The HMIcan include any interface that receives an input from the driver indicative of a desire of the driver to alter any parameter of the powertrainsuch as a torque output. In some examples, the HMI can be arranged on a steering wheel of the electrified vehicle.

While the vehicle control systemis shown as a single controller, it will be appreciated that more controllers and/or modules, such as a supervisory electrified vehicle control module, a battery control module, a motor control module and a chassis stability module, and various additional controllers described herein can be utilized to control various vehicle components of the hybrid electric vehicle. In this regard, various controllers and modules are configured to communicate with each other, utilizing different sensor inputsand calculated parameters as disclosed herein for controlling operation of the powertrain.

With additional reference now to, a plotof drive torque over time illustrating driver demand reference trajectory, traditional non-linear driver demand filteringand a model-based controlaccording to one example of the present disclosure. A lash zoneis defined as a backlash zone (space) between corresponding teeth of intermeshing gears in the powertrain (e.g., electric motors, etc.). Intermeshing gears in the powertrain can be any gears such as gear reduction gears and/or any gears between the electric motorand the driveline.

Traditional solutions to active damping, such as the traditional non-linear driver demand filteringinvolved two steps. The supervisory controller generated an optimal torque command using the minimum states of the system. This torque command underwent shaping through multiple filters at different parts of the system path. The input to the system, the driver demand, was shaped. The actuator commands, such as engine torque command and motor torque command, where shaped again as post-optimization outputs. Various types of shaping were applied, including one for controlling the powertrain at zero Newton meters (Mn) of torque or the lash zone within the transmission. Other shaping's controlled the system in its linear zone, where system stiffnesses could still be excited if not controlled appropriately. Of note, the traditional non-linear driver demand filteringincludes a small slope region(reduced torque increase over time) where the lash zoneis crossed (the corresponding intermeshing teeth rotate from a non-contacting position to a contacting position). When the increasing torque is slow and smooth, the lash is mitigated. However, with the traditional non-linear driver demand filtering, the torque response that the driver desires is delayed.

As shown, the model-based controlprovides a quick reaction (torque request is generated sooner compared to the traditional non-linear driver demand filtering), while still minimizing lash observed by the driver. With the system of the present disclosure, lash can be crossed much more quickly and climb in torque towards the driver intent within a time period shorter than the cascade of filters that are currently present in the traditional filtering.

With additional reference now to, a functional block diagram of a hybrid supervisory controlimplemented by the controllerof the hybrid electric vehicleofaccording to one example of the present disclosure will be described. The hybrid supervisory controlgenerally determines what torque is needed to deliver from a respective gear to the drive wheels. In examples, the hybrid supervisory controlcan operate in open loop to provide a motor torque command.

The hybrid supervisory controlincludes a static optimization module, a transmission model, a shift execution module, a multi-axle control vectoring module, a dynamic torque optimization moduleand a motor torque determination module.

The static optimization moduleemploys iterative algorithms to systematically explores all possible powertrain state combinations, seeking the optimal configuration. The dynamic torque optimization moduleiteratively adjusts torque commands to the ICEand the electric motorsto identify the most suitable combination of powertrain actuators. The combination achieves the desired driver demand while concurrently governing various powertrain inertias to adhere to their prescribed speed trajectories. The dynamic torque moduleoutputs a torque commandto the ICE. The motor torque determination moduleoutputs torque commandsto the motors. In effect, the hybrid supervisory controldevelops the driver demand reference trajectory(drive torque requestedby the driver to the drive wheels),.

is a functional block diagram of a supervisory motor torque controlimplemented by the controllerof the hybrid electric vehicleofaccording to one example of the present disclosure. The supervisory motor torque controlgenerally a supervisory torque determination moduleand a motor controller. The supervisory torque determination modulereceives a driver demand shaping output, and a reported engine torque signal. The driver demand shaping outputcan be generated from a driver demand shaping modulethat receives a signal from the accelerator pedal.

The supervisory torque determination moduleincludes a rigid driveline model (TSTF), an open loop electric motor torque calculation module, and an active damping controller. The active damping controllerincludes an optimal state feedback controllerand an optimal state estimator. The open loop electric motor torque calculationis explained above with respect to the hybrid supervisory controlof. The active damping controlleroutputs modified torque commandsassigned to the respective electric motors. The modified torque commandsare also represented by the model based control traceon. Further explanation of the operation of the active damping controlleris discussed below with respect to.

The MCPincludes an active electric motor dampening (AEMD) modulethat receives the torque commandsand outputs modified torque actuation signalsto the respective electric motors. The MCPprovides measurement feedbackto the active damping controller.

is a schematic illustration of a generic powertrainof which the active damping control can be applied to according to various examples of the present disclosure. The powertraingenerally includes a first motor, a second motor, a first axle, a second axleand a vehicle inertia. A first backlash non-linearityis disposed generally between the first electric motorand the vehicle inertia. A second backlash non-linearityis disposed generally between the second electric motorand the vehicle inertia. Both electric motors,are represented by an inertia JA and JB, respectively with torques Tand Tacting on each inertia. Both electric motor's position is represented by θand θ. Each motor shaft,entering backlash is denoted as θand θ. The vehicle's inertia is represented by Jwith Tacting on the vehicle's inertia. This torque represents the combination of friction brake torque and road load torque. The vehicles position is represented by θ. Axle shafts connecting electric motorsandto vehicle inertia are shown, having spring constant k and damping bi along with backlash angle βand βrespectively.

Each motor shaft,can be either in a contact state or within backlash. Contact states include both positive and negative contact, depending on the direction of the motor torque. This model can be represented in a time-varying state space form, thus having four distinct states: 1) all contact; 2) the first motorin contact with the second motorin backlash; 3) the first motorin backlash with the second motorin contact; and 4) both the first and second motors,in backlash.

The dynamic equations of the system are described below:

Consider backlash θand θwith size 2βand 2βas follows:

Now, θand θin the equations above can be replaced by equations including backlash:

From the above equations, one can see that two modes of operation exist for each shaft. A contact zone and backlash zone that is given by the following dynamics of the backlash state. Generalizing backlash dynamics for a shaft from above we get:

In order to design an optimal controller, states variables that are intuitive and have easy reference trajectories are considered, the selection of these states is a crucial consideration since reference trajectories for these states are essential for the effective operation of the controller.

In modeling a powertrain using a two mass-spring-damper system, the conventional approach found in the literature, typically includes the position and velocity of the masses as states of the system. Additionally, intermediate states of actuators, such as engines with slower dynamics, are considered. However, defining reference trajectories for position states becomes exceedingly challenging, particularly when supervisory controls dictate the system should follow a two mass-rigidly coupled model.

One novel contribution to this architecture, is how we addressed this challenge and ensured that the selected state variables have clear physical interpretations, we have replaced position states with their equivalent spring torque states. This replacement facilitates the construction of ‘intuitive’ Q (state error weight) and R (input effort weight) matrices used in optimal control to appropriately weigh each state. As a result, the state variables chosen for optimization across all proposed control methods are as follows:

The desired reference trajectories for these states are derived from the simplified reference model plant utilized by the supervisory controller's torque path. This essentially means that the supervisory control can continue to use a model with limit states, thus managing to keep the additional complexity of lash control and stiffness control outside optimization algorithms which would have required extremely high computational burden. In this reference model, motor speed outputs are readily available, and shaft torques are assumed to be equal to the torque transmitted by the rigid shaft connecting the motor actuators to vehicle inertia, or in other words, the output of supervisory control without the excessive filtering.